Maverick to gptoss120/scout nightmare

Hi all, I need someone who could assist with my prompt. It worked well in maverick but it was broken completely in gptoss120b and scout.

Gptoss120b worked well but has a lot of “half” responses example the bot would respond: this is a small one, I had one that the response was 40 lines.

Thank yo…

I appologize….

System error..

Ou

So now I have a few versions of my prompt but each having their own set of issues with either gptoss120 and llama4 scout.

Just let me know if you could assist, maybe over discord or so

broken completely in gptoss120b…

maybe modify the prompt?

Already done that.

I’m close to having it stable again, but still running into some issues. I’m still very new with prompting so it is taking me much longer to solve.

Yep, prompt engineering all the way.